gen - add 3D mixed support

cuda - fix 2D flattening

gpu - use 2d Flat variants in gen

gpu - add 2d Flattened variants of functions

gpu - isolate core 2D tensor logic to allow flat version

gpu - add P_1D to template args for AtPoints

hip - use BASIS_T_1D in codegen

cuda - use BASIS_T_1D in codegen

Add GetMinPointsInElement and GetMinMaxPointsInElement for elem restriction at points

gpu - add Transpose/TransposeAdd variants for AtPoints

gpu - fallback if nontensor shared uses too much mem

hip - fix missing template, compile values, fn names

gpu - fix AtPoints transpose shift

test - shrink sizes in t319 for non-tensor

hip - add nontensor shared

gpu - use gen LoadMatrix in shared

cuda - add nontensor shared

hip - AtPoints for hip/gen

minor - fix style in gen templates

gpu - minor fix to 1d AtPoints basis transpose

cuda - AtPoints for cuda/gen

doc - update docstrings for RstrType and FEType

shared - AtPoints template changes for gen

gen - fun name standardization

op - fix FLOPs estimates AtPoints